Preferences for association with co-ethnics can be expected among members of any ethnic group.37 These preferences will not necessar-ily be of positive value. In some circumstances people wish to avoid co-ethnics. Their preferences will then be of negative value. Because persons on both sides of a colour-based distinction tend to identify collectively, the possibility of their having co-ethnic neighbours may influence their search for housing and the possibility of their chil-dren having co-ethnic classmates may affect their choice of schools.
These are only preferences, so there will be a trade-off against other values.
A technique for measuring such preferences was developed for use in surveys in which samples of urban Malaysians have been
inter-viewed. Subjects were asked to predict how they thought others would decide in situations designed to measure preference for association with co-ethnics relative to the alternatives of personal gain and desire to meet what was felt as a personal obligation. In one situation they were told that Husin Ali, a representative Malay-Malaysian, bought his groceries from Ah Kow’s grocery shop, noted for its cheapness and close to his house. He had been told that someone called Ahmad was about to open a second grocery in the neighbourhood. Respondents were asked whether they thought that Husin Ali would transfer his custom to the new shop. Nothing in the interview said that Husin Ali was of Malay origin, or Ah Kow of Chinese origin. Those interviewed will have made this inference. In research elsewhere, the names or photographs of representative persons can be varied to discover more about the processes of social cognition.38
There was a common belief that Chinese-origin shopkeepers sold groceries more cheaply. Would Husin Ali prefer to help his co-ethnic (Ahmad), or would he buy where prices were lower (Ah Kow)? Were he to patronize Ahmad, this would be taken as an expression of social alignment based upon a preference for association with a co-ethnic.
The strength of such a preference could be measured, for example, by finding whether Husin Ali was predicted to continue shopping with Ahmad, if, other things being equal, his prices were 2, 4, 6 or 8 per cent higher. In a shopping situation, some individuals will have a preference of zero for association with a co-ethnic; others may have a higher preference, depending perhaps upon their personalities, their financial circumstances or the social pressures they experience.
A prediction that Husin Ali would prefer to shop with his co-ethnic could be seen as an estimation of his individual likes and dislikes, or as reflecting his solidarity with the co-ethnics who have made him the person he is. This latter aspect was measured in the research by asking respondents how they thought Husin Ali’s mother would wish him to act in the situations studied. The questions were varied to measure the preference for association with a co-ethnic by comparison with an expected financial gain, a gain in social status and the sense of obligation to a fellow employee. They were repeated in a study of the predicted ethnic preferences of a Chinese-Malaysian.
Just as many individuals will have a preference, in given situations, for association with a co-ethnic, so they may have preferences for as-sociation with someone of the same national origin, the same reli-gion, the same gender, the same social class or a speaker of the same
language. Questions could be devised that would enable an investiga-tor to measure the strength of one such association relative to others.
The components of forms of behaviour that have been aggregated as
‘racial’ can be separated. Methods of this kind provide better predic-tions of likely behaviour than the sorts of question posed in question-naire research.
Where do ethnic preferences come from? Every child is inducted into a pre-existing network of kin, neighbours and acquaintances from whom he or she can expect support. So preferences for association with certain kinds of other people exist from an early age. They persist only if they are reinforced by everyday experiences of reciprocity. It is on this basis that sets or groups of persons sharing similar values take shape. Andreas Wimmer has investigated the process by reanalysing information in the European Social Survey, a data set with more than 100,000 individuals and 380 ethnic groups in 24 countries. He found that only between 2 and 3 per cent of the variation in the measured values was located at the level of the ethnic group; between 7 and 16 per cent occurred at the level of the country, and between 80 and 90 per cent at the level of the individual. Individuals who reported Islam as their religion did not diverge from the basic values of their coun-tries any more than did Catholics.
The analysis showed that immigrants carried with them the value orientations into which they had been socialized, but that political exclusion had no effect on the value orientation of first-generation settlers. In the second generation, matters were very different. In the groups experiencing political exclusion, second-generation respon-dents deviated from the measured values three times more than their peers in groups that were not excluded. This supported the thesis that social closure increases the sharing of values in an excluded group.
Earlier in this book, it was held that many social relations are mul-tidimensional, in that someone who has been playing one role can switch to another role. Communities are characterized by the diver-sity of the bases on which members can interact with one another.
Wimmer takes this argument further by identifying four processes (he calls them mechanisms) by which different relationships can be tied with one another: availability, propinquity, homophily (or mem-bership sharing) and balancing processes. Their influence is mea-sured by use of a data set recording the social ties revealed by 1,640 students at a private college in the United States, 736 of whom posted photographs of their friends on Facebook.
This enabled Wimmer and his team to discover the existence of seventeen different communities, some ethnic, some based on the subjects studied and some tying together fans of particular bands or styles of music. The team concluded that ‘racial homophily does not represent the prime principle of tie formation … despite the emphasis on “race” that we find in many lay and sociological accounts of Amer-ican society’.39 The research opened up new techniques for comparing the strength of preferences for association with co-ethnics relative to other sentiments and ties governing the formation of social bonds.